%load_ext autoreload
%autoreload 2
%pylab inline

Populating the interactive namespace from numpy and matplotlib

%reset_selective -f  .*(?<!_)$
from custom_imports import *
from pathos.multiprocessing import ProcessingPool, ThreadingPool
#p = ProcessingPool(30)

masked_images_   = (joblib.load('cleaned_muscles.pkl'))
raw_images_ = (joblib.load('raw_images.pkl'))
original_images_ = (joblib.load('images.pkl'))
masked_images_   = [masked_images_[i] for i in A]
raw_images_ = [raw_images_[i] for i in A]
original_images_ = [original_images_[i] for i in A]
#masked_images   = bag.from_sequence(_masked_images)
#raw_images   = bag.from_sequence(_raw_images)
#original_images = bag.from_sequence(_original_images)
#del _masked_images  
#del _raw_images
#del _original_images

plt.imshow(_masked_images[0])

<matplotlib.image.AxesImage at 0x719e5260b358>

img_indx = np.random.randint(len(_raw_images))
scnd_img_to_drw = get_torso_outline((_original_images[img_indx], _masked_images[img_indx]-1))
if scnd_img_to_drw is None:
    print("is None")
else:
    plt.imshow(scnd_img_to_drw)

((1400, 1991, 3), (1400, 1991))

plt.imshow(_masked_images[img_indx])
plt.show()
plt.imshow(_raw_images[img_indx])
plt.show()
plt.imshow(scnd_img_to_drw)

<matplotlib.image.AxesImage at 0x719e52415908>

%reset_selective -f  .*(?<!_)$
from custom_imports import *

#PARAMS
dsize = resize_width, resize_height = 300, 300
neighSize = 41
STEP = 2

def gray_scale(x):
    if len(x.shape)>2:
        return x.mean(axis=2)
def min_1(x):
    return x-1   
x = p.map(gray_scale, _raw_images)
y = p.map(min_1, _masked_images)#.map(mark_margins, n_size=neighSize)

cond = p.map(get_shape_ratios, zip(x, y))

x = [i for i, c in zip(x, cond) if c]
y = [i for i, c in zip(y, cond) if c]
z = p.map(get_torso_outline, zip(x, y))

def mardas(i):
    return resize_image(i, dsize=dsize)
rszd_x = map(mardas, x)
rszd_x = list(rszd_x)
rszd_y = list(map(mardas, y))
rszd_z = list(map(lambda i: None if i is None else resize_image(i, dsize=dsize), z))
nrmlzd_rszd_x = list(map(normalize_image, rszd_x))
IXYZ = list((i, x, y, z) for i, x, y, z in zip(A, nrmlzd_rszd_x, rszd_y, rszd_z) if z is not None)
joblib.dump(IXYZ, 'IXYZ.pkl')

%reset_selective -f  .*(?<!_)$
from custom_imports import *
IXYZ = joblib.load('IXYZ.pkl')
i = np.random.randint(len(IXYZ))
fig, axes = plt.subplots(1, 3, figsize=(15, 5))
axes = axes.reshape(-1)
for i, x in enumerate(IXYZ[i][1:]):
    axes[i].imshow(x)

#preload = xy_.compute()

Preparing sampling

At this point, let's just take random samples, in the future I can change this so I can decide what proportion of each label I want to pick; that is, setting a weight for each label.

It is important to note that in order to make random kernel truly random, we need to explicitly pass the random seed to them. Otherwise the random states for all kernels (i.e. get_sample below) will have the same random states, leading to the same results on different runs.

Next Step

XY is saved in ./cleaned_data.pkl.

%reset_selective -f  .*(?<!_)$
%load_ext autoreload
%autoreload 2
from custom_imports import *
def get_neighbors_masked(ixyz, neigh_size):
    m = neigh_size//2
    print(m)
    i, x, y, z= ixyz
    # change later
    x = (x-x.mean())/x.std()
    w, h, *_ = x.shape
    X = []
    Y = []
    indices = list(zip(*np.where(z)))
    for i, j in indices:
        tmp = x[i-m:i+m, j-m:j+m]
        if tmp.shape == (2*m, 2*m):
            X.append(tmp)
            Y.append(y[i, j])
    return np.array(X), np.array(Y)

#PARAMS
dsize = resize_width, resize_height = 300, 300
neighSize = 41
STEP = 2
p = ProcessingPool(30)
IXYZ = joblib.load("IXYZ.pkl")
mardas = p.map(lambda x: get_neighbors_masked(x, neigh_size=neighSize), IXYZ)
X = np.concatenate([x[0] for x in mardas])
Y = np.concatenate([x[1] for x in mardas])

The autoreload extension is already loaded. To reload it, use:
  %reload_ext autoreload

X.shape, Y.shape
plt.imshow(X[156])

(11507474,)

Model v02

# Keras simple deep model:
def build_v02(width, height, depth, reg, init="he_normal"):
    # initialize the model along with the input shape to be
    # "channels last" and the channels dimension itself
    model = Sequential()
    inputShape = (height, width, depth)
    chanDim = -1

    # if we are using "channels first", update the input shape
    # and channels dimension
    if K.image_data_format() == "channels_first":
        inputShape = (depth, height, width)
        chanDim = 1
        
    # our first CONV layer will learn a total of 16 filters, each
    # Of which are 7x7 -- we'll then apply 2x2 strides to reduce
    # the spatial dimensions of the volume
    #model.add(Conv2D(32, (10, 10), strides=(1, 1), padding="same",
    #    kernel_initializer=init,# kernel_regularizer=reg,
    #    input_shape=inputShape))

    # here we stack two CONV layers on top of each other where
    # each layerswill learn a total of 32 (3x3) filters
    model.add(Conv2D(32, (10, 10), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    model.add(Activation("relu"))
    #model.add(BatchNormalization(axis=chanDim))
    model.add(Conv2D(32, (5, 5), strides=(1, 1), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    model.add(Activation("relu"))
    #model.add(BatchNormalization(axis=chanDim))
    model.add(Dropout(0.25))
    # stack two more CONV layers, keeping the size of each filter
    # as 3x3 but increasing to 64 total learned filters
    #model.add(Conv2D(64, (3, 3), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    #model.add(Conv2D(64, (3, 3), strides=(1, 1), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    #model.add(Dropout(0.25))
#
    ## increase the number of filters again, this time to 128
    #model.add(Conv2D(128, (3, 3), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    model.add(Conv2D(10, (3, 3), strides=(1, 1), padding="same",
        kernel_initializer=init,  #kernel_regularizer=reg
                    )
             )
    model.add(Activation("relu"))
    model.add(Conv2D(1, (3, 3), strides=(1, 1), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    model.add(Activation("relu"))
    model.add(Flatten())
    model.add(Dense(100))
    model.add(Dense(10))
    model.add(Dense(1))
    model.add(Activation("sigmoid"))
    #model.add(BatchNormalization(axis=chanDim))
    return model

Model v03

# Keras simple deep model:
def build_v03(width, height, depth, reg, init="he_normal"):
    # initialize the model along with the input shape to be
    # "channels last" and the channels dimension itself
    model = Sequential()
    inputShape = (height, width, depth)
    chanDim = -1

    # if we are using "channels first", update the input shape
    # and channels dimension
    if K.image_data_format() == "channels_first":
        inputShape = (depth, height, width)
        chanDim = 1
        
    # our first CONV layer will learn a total of 16 filters, each
    # Of which are 7x7 -- we'll then apply 2x2 strides to reduce
    # the spatial dimensions of the volume
    model.add(Conv2D(32, (30, 30), strides=(1, 1), padding="same",
        kernel_initializer=init,# kernel_regularizer=reg,
        input_shape=inputShape))
   
    model.add(Conv2D(32, (15, 15), strides=(1, 1), padding="same",
        kernel_initializer=init,# kernel_regularizer=reg,
        input_shape=inputShape))

    # here we stack two CONV layers on top of each other where
    # each layerswill learn a total of 32 (3x3) filters
    model.add(Conv2D(32, (10, 10), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    #model.add(Activation("relu"))
    #model.add(BatchNormalization(axis=chanDim))
    model.add(Conv2D(32, (5, 5), strides=(1, 1), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    #model.add(Activation("relu"))
    #model.add(BatchNormalization(axis=chanDim))
    #model.add(Dropout(0.25))
    # stack two more CONV layers, keeping the size of each filter
    # as 3x3 but increasing to 64 total learned filters
    #model.add(Conv2D(64, (3, 3), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    #model.add(Conv2D(64, (3, 3), strides=(1, 1), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    #model.add(Dropout(0.25))
#
    ## increase the number of filters again, this time to 128
    #model.add(Conv2D(128, (3, 3), padding="same",
    #    kernel_initializer=init, kernel_regularizer=reg))
    #model.add(Activation("relu"))
    ##model.add(BatchNormalization(axis=chanDim))
    model.add(Conv2D(10, (3, 3), strides=(1, 1), padding="same",
        kernel_initializer=init,  #kernel_regularizer=reg
                    )
             )
    model.add(Activation("relu"))
    model.add(Conv2D(1, (3, 3), strides=(1, 1), padding="same",
        kernel_initializer=init, #kernel_regularizer=reg)
                    ))
    #model.add(Activation("relu"))
    #model.add(Flatten())
    #model.add(Dense(100))
    #model.add(Dense(10))
    #model.add(Dense(1))
    model.add(Activation("sigmoid"))
    #model.add(BatchNormalization(axis=chanDim))
    return model

*_, m, n = X.shape
#if len(_) == 1:
#    X = X.reshape(-1, m, n, 1)
#    Y = Y.reshape(-1, 1)

Xtr, Xts, Ytr, Yts = train_test_split(X.reshape(*X.shape, 1), Y.reshape(*Y.shape, 1)==1, test_size=.2)

# creating and starting keras model example:
opt = Adam(lr=1e-4)#, decay=1e-4 / epochs)
#v02
#model = build(width=neighSize, height=neighSize, depth=1, reg=l2(0.0005))
model = build_v02(width=m, height=n, depth=1, reg=l2(0.0005))
model.compile(loss=losses.binary_crossentropy, optimizer=opt, metrics=["accuracy"])

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-50-4152538a361a> in <module>
      4 #v02
      5 #model = build(width=neighSize, height=neighSize, depth=1, reg=l2(0.0005))
----> 6 model = build_v02(width=m, height=n, depth=1, reg=l2(0.0005))
      7 model.compile(loss=losses.binary_crossentropy, optimizer=opt, metrics=["accuracy"])

NameError: name 'build_v02' is not defined

1 - (Ytr.sum()/(Ytr>-1).sum())

0.6613350953766025

Xtr.shape

(9205979, 40, 40, 1)

#model.fit(X, Y, batch_size=1)
# verion02
#model.fit_generator(gen(Xtr, Ytr, 10, neighSize), 1000, epochs=1000)#, validation_data=gen(Xts, Yts, 10, neighSize))
model.fit(Xtr, Ytr, batch_size=800, epochs=500)#, validation_data=gen(Xts, Yts, 10, neighSize))

Epoch 1/500
9205979/9205979 [==============================] - 3418s 371us/step - loss: 0.1841 - acc: 0.9273
Epoch 2/500
4750400/9205979 [==============>...............] - ETA: 27:30 - loss: 0.1428 - acc: 0.9442

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3416s 371us/step - loss: 0.1381 - acc: 0.9461
Epoch 3/500
 156000/9205979 [..............................] - ETA: 57:26 - loss: 0.1288 - acc: 0.9497

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

4910400/9205979 [===============>..............] - ETA: 26:36 - loss: 0.1263 - acc: 0.9507

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3416s 371us/step - loss: 0.1239 - acc: 0.9516
Epoch 4/500
3279200/9205979 [=========>....................] - ETA: 36:36 - loss: 0.1177 - acc: 0.9541

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3411s 371us/step - loss: 0.1156 - acc: 0.9548
Epoch 5/500
1606400/9205979 [====>.........................] - ETA: 46:56 - loss: 0.1116 - acc: 0.9564

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

6348800/9205979 [===================>..........] - ETA: 17:38 - loss: 0.1105 - acc: 0.9567

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3408s 370us/step - loss: 0.1098 - acc: 0.9570
Epoch 6/500
1877600/9205979 [=====>........................] - ETA: 45:14 - loss: 0.1066 - acc: 0.9584

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

6419200/9205979 [===================>..........] - ETA: 17:11 - loss: 0.1060 - acc: 0.9586

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3407s 370us/step - loss: 0.1056 - acc: 0.9587
Epoch 7/500
1952800/9205979 [=====>........................] - ETA: 44:45 - loss: 0.1032 - acc: 0.9595

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

6703200/9205979 [====================>.........] - ETA: 15:26 - loss: 0.1027 - acc: 0.9598

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3406s 370us/step - loss: 0.1024 - acc: 0.9599
Epoch 8/500
2244800/9205979 [======>.......................] - ETA: 42:56 - loss: 0.1005 - acc: 0.9606

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

6988800/9205979 [=====================>........] - ETA: 13:40 - loss: 0.1000 - acc: 0.9608

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

2528800/9205979 [=======>......................] - ETA: 41:11 - loss: 0.0984 - acc: 0.9614

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7011200/9205979 [=====================>........] - ETA: 13:31 - loss: 0.0978 - acc: 0.9617

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

2546400/9205979 [=======>......................] - ETA: 41:05 - loss: 0.0962 - acc: 0.9622

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7284000/9205979 [======================>.......] - ETA: 11:51 - loss: 0.0959 - acc: 0.9624

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

2788800/9205979 [========>.....................] - ETA: 39:35 - loss: 0.0945 - acc: 0.9629

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7203200/9205979 [======================>.......] - ETA: 12:20 - loss: 0.0944 - acc: 0.9629

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

2708800/9205979 [=======>......................] - ETA: 40:03 - loss: 0.0933 - acc: 0.9633

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7416800/9205979 [=======================>......] - ETA: 11:01 - loss: 0.0931 - acc: 0.9634

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

2923200/9205979 [========>.....................] - ETA: 38:44 - loss: 0.0921 - acc: 0.9637

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7633600/9205979 [=======================>......] - ETA: 9:41 - loss: 0.0918 - acc: 0.9638

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3136800/9205979 [=========>....................] - ETA: 37:25 - loss: 0.0908 - acc: 0.9643

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7867200/9205979 [========================>.....] - ETA: 8:15 - loss: 0.0907 - acc: 0.9643

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3401600/9205979 [==========>...................] - ETA: 35:46 - loss: 0.0900 - acc: 0.9645

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

8112800/9205979 [=========================>....] - ETA: 6:44 - loss: 0.0898 - acc: 0.9646

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3618400/9205979 [==========>...................] - ETA: 34:26 - loss: 0.0891 - acc: 0.9648

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

9205979/9205979 [==============================] - 3404s 370us/step - loss: 0.0889 - acc: 0.9649
Epoch 17/500
9205979/9205979 [==============================] - 3405s 370us/step - loss: 0.0880 - acc: 0.9653
Epoch 18/500
9205979/9205979 [==============================] - 3403s 370us/step - loss: 0.0873 - acc: 0.9655
Epoch 19/500
2898400/9205979 [========>.....................] - ETA: 38:52 - loss: 0.0867 - acc: 0.9657

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7619200/9205979 [=======================>......] - ETA: 9:46 - loss: 0.0866 - acc: 0.9657

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3031200/9205979 [========>.....................] - ETA: 38:03 - loss: 0.0862 - acc: 0.9659

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7768000/9205979 [========================>.....] - ETA: 8:51 - loss: 0.0860 - acc: 0.9659

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3169600/9205979 [=========>....................] - ETA: 37:12 - loss: 0.0853 - acc: 0.9662

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

7913600/9205979 [========================>.....] - ETA: 7:57 - loss: 0.0854 - acc: 0.9662

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3447200/9205979 [==========>...................] - ETA: 35:29 - loss: 0.0849 - acc: 0.9664

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

8185600/9205979 [=========================>....] - ETA: 6:17 - loss: 0.0849 - acc: 0.9664

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3718400/9205979 [===========>..................] - ETA: 33:49 - loss: 0.0844 - acc: 0.9666

IOPub message rate exceeded.
The notebook server will temporarily stop sending output
to the client in order to avoid crashing it.
To change this limit, set the config variable
`--NotebookApp.iopub_msg_rate_limit`.

Current values:
NotebookApp.iopub_msg_rate_limit=1000.0 (msgs/sec)
NotebookApp.rate_limit_window=3.0 (secs)

3903200/9205979 [===========>..................] - ETA: 32:41 - loss: 0.0844 - acc: 0.9666

---------------------------------------------------------------------------
KeyboardInterrupt                         Traceback (most recent call last)
<ipython-input-89-48847dc380c3> in <module>
      2 # verion02
      3 #model.fit_generator(gen(Xtr, Ytr, 10, neighSize), 1000, epochs=1000)#, validation_data=gen(Xts, Yts, 10, neighSize))
----> 4 model.fit(Xtr, Ytr, batch_size=800, epochs=500)#, validation_data=gen(Xts, Yts, 10, neighSize))

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/keras/engine/training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)
   1035                                         initial_epoch=initial_epoch,
   1036                                         steps_per_epoch=steps_per_epoch,
-> 1037                                         validation_steps=validation_steps)
   1038 
   1039     def evaluate(self, x=None, y=None,

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/keras/engine/training_arrays.py in fit_loop(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)
    197                     ins_batch[i] = ins_batch[i].toarray()
    198 
--> 199                 outs = f(ins_batch)
    200                 outs = to_list(outs)
    201                 for l, o in zip(out_labels, outs):

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py in __call__(self, inputs)
   2670                     'In order to feed symbolic tensors to a Keras model '
   2671                     'in TensorFlow, you need tensorflow 1.8 or higher.')
-> 2672             return self._legacy_call(inputs)
   2673 
   2674 

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py in _legacy_call(self, inputs)
   2652         session = get_session()
   2653         updated = session.run(fetches=fetches, feed_dict=feed_dict,
-> 2654                               **self.session_kwargs)
   2655         return updated[:len(self.outputs)]
   2656 

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/client/session.py in run(self, fetches, feed_dict, options, run_metadata)
    787     try:
    788       result = self._run(None, fetches, feed_dict, options_ptr,
--> 789                          run_metadata_ptr)
    790       if run_metadata:
    791         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/client/session.py in _run(self, handle, fetches, feed_dict, options, run_metadata)
    995     if final_fetches or final_targets:
    996       results = self._do_run(handle, final_targets, final_fetches,
--> 997                              feed_dict_string, options, run_metadata)
    998     else:
    999       results = []

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/client/session.py in _do_run(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)
   1130     if handle is None:
   1131       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,
-> 1132                            target_list, options, run_metadata)
   1133     else:
   1134       return self._do_call(_prun_fn, self._session, handle, feed_dict,

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/client/session.py in _do_call(self, fn, *args)
   1137   def _do_call(self, fn, *args):
   1138     try:
-> 1139       return fn(*args)
   1140     except errors.OpError as e:
   1141       message = compat.as_text(e.message)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/client/session.py in _run_fn(session, feed_dict, fetch_list, target_list, options, run_metadata)
   1119         return tf_session.TF_Run(session, options,
   1120                                  feed_dict, fetch_list, target_list,
-> 1121                                  status, run_metadata)
   1122 
   1123     def _prun_fn(session, handle, feed_dict, fetch_list):

KeyboardInterrupt: 

import datetime, re
d = re.sub(r':| |\.', '_', f"{datetime.datetime.now()}")
model.save(f'best_so_far_neighbor_{neighSize}_{d}.model_v03')

ls *v03

best_so_far_neighbor_17_2020-04-13_17_23_17_464782.model_v03
best_so_far_neighbor_41_2020-04-14_14_40_34_896919.model_v03

from keras.models import load_model
model = load_model(f'best_so_far_neighbor_41_2020-04-14_14_40_34_896919.model_v03')

Using TensorFlow backend.

from functools import reduce
def predict_image(x, model, neigh_size):
    m = neigh_size//2
    xn = get_neighbors(x, neigh_size, step=1)
    print(x.shape)
    m = neigh_size//2
    *_, w, h = tuple(x-2*m for x in x.shape)
    yn = model.predict(xn.reshape(*xn.shape, 1))
    print(w, h, w*h, yn.shape)
    return yn.reshape(w, h)

n = np.random.randint(len(XYZ))
print(n)
res = predict_image(XYZ[n][0], model, neighSize)

62
(300, 300)
260 260 67600 (67600, 1)

m = neighSize//2
x = res>.1
fig, axes = plt.subplots(1, 3, figsize=(15, 5))
#axes[0].imshow(x)
axes[0].imshow(XYZ[n][0])
axes[1].imshow(XYZ[n][1][m:-m, m:-m])
#axes[3].imshow(XYZ[n][2][m:-m, m:-m])
mkkernel = lambda k: np.ones((k, k))
t = cv2.erode((XYZ[n][2][m:-m, m:-m])*x, mkkernel(2), 1)
axes[2].imshow(t)

<matplotlib.image.AxesImage at 0x793853cd0e10>

import datetime, re
d = re.sub(r':| |\.', '_', f"{datetime.datetime.now()}")
model2.save(f'best_so_far_neighbor_{neighSize}_{d}.model_v04')

ls *v03

best_so_far_neighbor_17_2020-04-13_17_23_17_464782.model_v03
best_so_far_neighbor_41_2020-04-14_14_40_34_896919.model_v03

from keras.models import load_model
model2 = load_model(f'best_so_far_neighbor_41_2020-04-14_14_40_34_896919.model_v03')

tf.test.is_gpu_available()

False

n = np.random.randint(len(XYZ))
print(n)
res = predict_image(XYZ[n][0], model2, neighSize)

55
(300, 300)
260 260 67600 (67600, 1)

res.shape

(260, 260)

XYZ[n][2].shape

(300, 300)

m = neighSize//2
x = res>.1
fig, axes = plt.subplots(1, 3, figsize=(15, 5))
#axes[0].imshow(x)
axes[0].imshow(XYZ[n][0])
axes[1].imshow(XYZ[n][1][m:-m, m:-m])
#axes[3].imshow(nw_XYZ[n][2][m:-m, m:-m])
mkkernel = lambda k: np.ones((k, k))
t = cv2.erode((XYZ[n][2][m:-m, m:-m])*x, mkkernel(2), 1)
axes[2].imshow(t)

<matplotlib.image.AxesImage at 0x7938400a0a58>

#np.random.choice(A, replace=False, size=10)
indices = np.random.choice(range(len(_masked_images)), size=15)
indices

array([136, 183, 392, 209, 278,  75, 482, 458, 441,  86, 226, 379, 269,
       124, 435])

mkdir 'masked'

mkdir: cannot create directory ‘masked’: File exists

for i in indices:
    cv2.imwrite('./masked/masked_%03d.png'%i, (_masked_images[i]-1)*255)

masked_021.png  masked_088.png  masked_144.png  masked_219.png  masked_289.png
masked_030.png  masked_120.png  masked_158.png  masked_239.png  masked_334.png
masked_062.png  masked_140.png  masked_195.png  masked_262.png  masked_436.png

import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
import os
import cv2
from dask.distributed import Client, LocalCluster
from dask import delayed, compute
from dask import bag, array as da
%load_ext autoreload
%autoreload 2
%pylab inline

/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:458: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint8 = np.dtype([("qint8", np.int8, 1)])
/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:459: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:460: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint16 = np.dtype([("qint16", np.int16, 1)])
/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:461: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:462: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint32 = np.dtype([("qint32", np.int32, 1)])
/home/bzr0014/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:465: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  np_resource = np.dtype([("resource", np.ubyte, 1)])

Populating the interactive namespace from numpy and matplotlib

# keras important imports:
# import the necessary packages
from keras.models import Sequential
from keras.models import load_model
from keras.layers.normalization import BatchNormalization
from keras.layers.convolutional import Conv2D
from keras.layers.core import Activation
from keras.layers.core import Flatten
from keras.layers.core import Dropout
from keras.layers.core import Dense
from keras import backend as K

from sklearn.preprocessing import LabelBinarizer
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report
from keras.preprocessing.image import ImageDataGenerator
from keras.optimizers import Adam
from keras.regularizers import l2

import glob
files = glob.glob('masked/*png')

import re

k = 10
def mardas(x):
    x = cv2.resize(x[:, :, 2], (300, 300), interpolation=cv2.INTER_NEAREST)
    y = cv2.dilate((x).astype('uint8'), np.ones((3,3)))
    indcs =  np.where(y>0)
    t = np.array(indcs).T
    f = np.array([(y>0)[i-k:i+k, j-k:j+k].reshape(-1) for i,j in zip(*indcs)])
    return y, indcs, np.concatenate([f, t], axis=1), y[indcs]

def mardas(args):
    x, y = args
    y = y.mean(axis=2)
    x = cv2.resize(x[:, :, 2], (300, 300), interpolation=cv2.INTER_NEAREST)
    y = cv2.resize(y, (300, 300))
    z = cv2.dilate((x).astype('uint8'), np.ones((3,3)))
    indcs =  np.where(z>0)
    t = np.array(indcs).T
    f = np.array([(z>0)[i-k:i+k, j-k:j+k] for i,j in zip(*indcs)])
    ff = np.array([y[i-k:i+k, j-k:j+k] for i,j in zip(*indcs)])
    f = f.reshape(*f.shape, 1)
    ff = ff.reshape(*ff.shape, 1)
    return z, indcs, (f, ff, t), z[indcs]

X_ = [_raw_images[int(re.findall(r'_(\d{3})\.', x)[0])] for x in files]
Xy = list(map(mardas, zip(map(cv2.imread, files), X_)))

from sklearn.preprocessing import OneHotEncoder, MinMaxScaler
encoder = OneHotEncoder()
scaler = MinMaxScaler()
#X = np.concatenate([x[2] for x in Xy])
#X = scaler.fit_transform(X)
X1 = np.concatenate([x[2][0] for x in Xy])
mu1, sig1 = X1.mean(), X1.std()
X1 = (X1-mu1)/sig1
X2 = np.concatenate([x[2][1] for x in Xy])
mu2, sig2 = X2.mean(), X2.std()
X2 = (X2-mu2)/sig2
X3 = np.concatenate([x[2][2] for x in Xy])
X3 = scaler.fit_transform(X3)
y = np.concatenate([x[3] for x in Xy]).reshape(-1 ,1)
y = encoder.fit_transform(y).toarray()

X1.shape, X2.shape, X3.shape, y.shape

((229227, 20, 20, 1), (229227, 20, 20, 1), (229227, 2), (229227, 11))

from sklearn.model_selection import train_test_split

rnd = np.random.rand(len(X1))
X1tr = X1[rnd>.1]
X1ts = X1[rnd<.1]
X2tr = X2[rnd>.1]
X2ts = X2[rnd<.1]
X3tr = X3[rnd>.1]
X3ts = X3[rnd<.1]
ytr = y[rnd>.1]
yts = y[rnd<.1]

def baseline_model():
	# create model
	model = Sequential()
	model.add(Dense(20, input_dim=(k*2)**2+2, activation='relu'))
	model.add(Dense(11, activation='sigmoid'))
	model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
	return model
model = baseline_model()

import keras
input1 = keras.layers.Input(shape=(2*k,2*k,1))
x1 = keras.layers.Conv2D(8, (3, 3), activation='relu', padding='same')(input1)
x1 = keras.layers.Conv2D(8, (3, 3), activation='relu', padding='same')(x1)
x1 = keras.layers.Flatten()(x1)
x1 = keras.layers.Dense(11, activation='relu')(x1)

input2 = keras.layers.Input(shape=(2*k,2*k,1))
x2 = keras.layers.Conv2D(8, (3, 3), activation='relu', padding='same')(input2)
x2 = keras.layers.Conv2D(8, (3, 3), activation='relu', padding='same')(x2)
x2 = keras.layers.Flatten()(x2)
x2 = keras.layers.Dense(11, activation='relu')(x2)


input3 = keras.layers.Input(shape=(2,))
x3 = keras.layers.Dense(11, activation='relu')(input3)
#x2 = keras.layers.Dense(11, activation='relu')(x2)
# equivalent to added = keras.layers.add([x1, x2])
added = keras.layers.Add()([x1, x2, x3])
added = keras.layers.Dense(11, activation='relu')(added)
out = keras.layers.Dense(11, activation='sigmoid')(added)
model = keras.models.Model(inputs=[input1, input2, input3], outputs=out)
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

model.fit([X1tr, X2tr, X3tr], ytr, batch_size=800, epochs=100, validation_data=([X1ts, X2ts, X3ts], yts))

nn = np.random.randint(len(Xy))
print(nn)
_1, _2, _3 = Xy[nn][:3]

13

_4, _5, _6 = (_3[0]-mu1)/sig1, (_3[1]-mu2)/sig2, scaler.transform(_3[2])
_ = model.predict([_4, _5, _6])
_ = _ / _.sum(axis=1, keepdims=1)
_ = encoder.inverse_transform(_>.1)

imt = np.zeros_like(_1)
#imt[_2] = _.reshape(-1)
imt[_2] = _.reshape(-1)

a = np.zeros_like(imt)
for i in np.where(np.bincount(imt.reshape(-1)))[0][1:]:
    tmp = cv2.erode((imt == i).astype('uint8'), np.ones((2, 2)))
    _, cnts, _ = cv2.findContours(tmp.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
    cnts = sorted(cnts, key=lambda x: -len(x))[:1]
    tmp = np.zeros_like(tmp)
    a += cv2.drawContours(tmp, cnts, -1, int(i), -1)
    #a   += cv2.dilate(tmp, np.ones((4, 4)))*i
plt.imshow(a)

<matplotlib.image.AxesImage at 0x793831965e48>

def mardas(x):
    imt = np.zeros(tuple(i+2*k for i in x.shape))
    imt[k:-k, k:-k] = x   
    x = imt
    indcs =  np.where(x>0)
    t = np.array(indcs).T
    f = np.array([(x>0)[i-k:i+k, j-k:j+k] for i,j in zip(*indcs)])
    f = f.reshape(*f.shape, 1)
    return x, indcs, [f, t], x[indcs]
def get_classes(x):
    _1, _2, (_31, _32), *_ = mardas(x)
    _4 = model.predict([(_31-mu)/std, scaler.transform(_32)])
    imt = np.zeros_like(x)
    yprd = encoder.inverse_transform(((_4 / _4.sum(axis=1, keepdims=True)) > .1)*1)
    imt[_2] = yprd.reshape(-1)
    return imt
plt.imshow(get_classes(_1))

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-357-d9e280a24025> in <module>
     14     imt[_2] = yprd.reshape(-1)
     15     return imt
---> 16 plt.imshow(get_classes(_1))

<ipython-input-357-d9e280a24025> in get_classes(x)
      9 def get_classes(x):
     10     _1, _2, _3, *_ = mardas(x)
---> 11     _4 = model.predict(scaler.transform(_3))
     12     imt = np.zeros_like(x)
     13     yprd = encoder.inverse_transform(((_4 / _4.sum(axis=1, keepdims=True)) > .1)*1)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/sklearn/preprocessing/_data.py in transform(self, X)
    412                         force_all_finite="allow-nan")
    413 
--> 414         X *= self.scale_
    415         X += self.min_
    416         return X

ValueError: operands could not be broadcast together with shapes (19343,198) (2,) (19343,198) 

plt.imshow(get_classes((res>.1) & XYZ[55][2][m:-m, m:-m]))

<matplotlib.image.AxesImage at 0x79383dc35470>

plt.imshow(cv2.dilate((_1>0).astype('uint8'), np.ones((3,3))))

<matplotlib.image.AxesImage at 0x79383ec710b8>

sss.max()

1

N = 10
l = int(np.log(N)/np.log(2))+1

2**(l+4) / N

25.6

t = np.arange(0, 2**(l+4), 25)
["{0:08b}".format(x) for x in t]

['00000000',
 '00011001',
 '00110010',
 '01001011',
 '01100100',
 '01111101',
 '10010110',
 '10101111',
 '11001000',
 '11100001',
 '11111010']

np.abs(t.reshape(1, -1) - np.random.randint(256, size=50).reshape(-1, 1)).argmin(axis=1)

array([ 2,  7,  6,  9,  7,  0,  5,  2,  2,  1,  8,  0,  8,  5,  3,  9,  8,
        7,  7,  8,  3,  2,  8,  9,  6,  4,  9,  5,  6,  0,  6,  9,  1,  1,
        1,  1,  2,  5,  0,  7,  9, 10,  4,  5,  8,  1,  6,  2, 10,  4])

a = np.random.rand(10, 3)

(1+a.dot(a.T**.2))#.sum(axis=0)

array([[2.56706585, 2.33048033, 2.57231358, 2.32754517, 2.39959548,
        2.46850553, 2.60619466, 2.59029428, 2.67219455, 2.50950931],
       [1.8461522 , 1.78965218, 1.81566946, 1.70656476, 1.82201851,
        1.77256526, 1.88060444, 1.90706822, 1.93438748, 1.87899436],
       [2.64719657, 2.34152327, 2.67964964, 2.41197942, 2.43244496,
        2.57600137, 2.67647735, 2.63300826, 2.74030125, 2.52457813],
       [1.71807919, 1.59335194, 1.72852871, 1.62273164, 1.6468314 ,
        1.70085784, 1.73123753, 1.71777445, 1.77027547, 1.66119075],
       [2.12460112, 2.04864037, 2.08534598, 1.9639864 , 2.13420411,
        2.07488206, 2.16669322, 2.20523974, 2.2645669 , 2.13484746],
       [2.30599632, 2.05685682, 2.33585305, 2.14819724, 2.17693592,
        2.30499027, 2.32398922, 2.29043869, 2.40263073, 2.16620148],
       [2.78404888, 2.54116125, 2.7774552 , 2.50184767, 2.60822343,
        2.65326233, 2.83435754, 2.82839538, 2.91010587, 2.74977171],
       [2.80534647, 2.62472982, 2.76827879, 2.51271159, 2.69203319,
        2.6588386 , 2.8681132 , 2.8944954 , 2.96385161, 2.8260797 ],
       [3.25670011, 2.9838767 , 3.23330169, 2.92483324, 3.117833  ,
        3.13943235, 3.32248598, 3.33648274, 3.45933851, 3.20465454],
       [2.57840033, 2.43734777, 2.53707565, 2.29010902, 2.44356081,
        2.38755309, 2.64071151, 2.66746727, 2.696476  , 2.65116679]])

(1+a.dot(a[0].T**.2))#.sum()

array([2.56706585, 1.8461522 , 2.64719657, 1.71807919, 2.12460112,
       2.30599632, 2.78404888, 2.80534647, 3.25670011, 2.57840033])

import datetime, re
d = re.sub(r':| |\.', '_', f"{datetime.datetime.now()}")
model.save(f'best_classification_{neighSize}_{d}.model_v03')

#ress = list(map(lambda x: predict_image(x[0], model2, neighSize), XYZ))

(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)
(300, 300)
260 260 67600 (67600, 1)

XYZT = [(x,y,z,t) for (x,y,z), t in zip(XYZ, ress)]

len(XYZT[0])

4

rs = list(map(lambda x: x[2][m:-m, m:-m] * x[3], XYZT))

for x, y, z, t in XYZT:
    fig, axes = plt.subplots(1, 3)
    axes[0].imshow(x)
    axes[1].imshow(z[m:-m, m:-m] * t)
    axes[2].imshow(y)
    plt.show()

---------------------------------------------------------------------------
KeyboardInterrupt                         Traceback (most recent call last)
<ipython-input-614-6ba8ef6d773c> in <module>
      4     axes[1].imshow(z[m:-m, m:-m] * t)
      5     axes[2].imshow(y)
----> 6     plt.show()

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/pyplot.py in show(*args, **kw)
    267     """
    268     global _show
--> 269     return _show(*args, **kw)
    270 
    271 

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/ipykernel/pylab/backend_inline.py in show(close, block)
     41             display(
     42                 figure_manager.canvas.figure,
---> 43                 metadata=_fetch_figure_metadata(figure_manager.canvas.figure)
     44             )
     45     finally:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/display.py in display(include, exclude, metadata, transient, display_id, *objs, **kwargs)
    311             publish_display_data(data=obj, metadata=metadata, **kwargs)
    312         else:
--> 313             format_dict, md_dict = format(obj, include=include, exclude=exclude)
    314             if not format_dict:
    315                 # nothing to display (e.g. _ipython_display_ took over)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/formatters.py in format(self, obj, include, exclude)
    178             md = None
    179             try:
--> 180                 data = formatter(obj)
    181             except:
    182                 # FIXME: log the exception

<decorator-gen-9> in __call__(self, obj)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/formatters.py in catch_format_error(method, self, *args, **kwargs)
    222     """show traceback on failed format call"""
    223     try:
--> 224         r = method(self, *args, **kwargs)
    225     except NotImplementedError:
    226         # don't warn on NotImplementedErrors

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/formatters.py in __call__(self, obj)
    339                 pass
    340             else:
--> 341                 return printer(obj)
    342             # Finally look for special method names
    343             method = get_real_method(obj, self.print_method)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/pylabtools.py in <lambda>(fig)
    246 
    247     if 'png' in formats:
--> 248         png_formatter.for_type(Figure, lambda fig: print_figure(fig, 'png', **kwargs))
    249     if 'retina' in formats or 'png2x' in formats:
    250         png_formatter.for_type(Figure, lambda fig: retina_figure(fig, **kwargs))

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/IPython/core/pylabtools.py in print_figure(fig, fmt, bbox_inches, **kwargs)
    130         FigureCanvasBase(fig)
    131 
--> 132     fig.canvas.print_figure(bytes_io, **kw)
    133     data = bytes_io.getvalue()
    134     if fmt == 'svg':

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/backend_bases.py in print_figure(self, filename, dpi, facecolor, edgecolor, orientation, format, bbox_inches, **kwargs)
   2089                     orientation=orientation,
   2090                     bbox_inches_restore=_bbox_inches_restore,
-> 2091                     **kwargs)
   2092             finally:
   2093                 if bbox_inches and restore_bbox:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/backends/backend_agg.py in print_png(self, filename_or_obj, metadata, pil_kwargs, *args, **kwargs)
    525 
    526         else:
--> 527             FigureCanvasAgg.draw(self)
    528             renderer = self.get_renderer()
    529             with cbook._setattr_cm(renderer, dpi=self.figure.dpi), \

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/backends/backend_agg.py in draw(self)
    386         self.renderer = self.get_renderer(cleared=True)
    387         with RendererAgg.lock:
--> 388             self.figure.draw(self.renderer)
    389             # A GUI class may be need to update a window using this draw, so
    390             # don't forget to call the superclass.

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     36                 renderer.start_filter()
     37 
---> 38             return draw(artist, renderer, *args, **kwargs)
     39         finally:
     40             if artist.get_agg_filter() is not None:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/figure.py in draw(self, renderer)
   1707             self.patch.draw(renderer)
   1708             mimage._draw_list_compositing_images(
-> 1709                 renderer, self, artists, self.suppressComposite)
   1710 
   1711             renderer.close_group('figure')

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/image.py in _draw_list_compositing_images(renderer, parent, artists, suppress_composite)
    133     if not_composite or not has_images:
    134         for a in artists:
--> 135             a.draw(renderer)
    136     else:
    137         # Composite any adjacent images together

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     36                 renderer.start_filter()
     37 
---> 38             return draw(artist, renderer, *args, **kwargs)
     39         finally:
     40             if artist.get_agg_filter() is not None:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/axes/_base.py in draw(self, renderer, inframe)
   2645             renderer.stop_rasterizing()
   2646 
-> 2647         mimage._draw_list_compositing_images(renderer, self, artists)
   2648 
   2649         renderer.close_group('axes')

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/image.py in _draw_list_compositing_images(renderer, parent, artists, suppress_composite)
    133     if not_composite or not has_images:
    134         for a in artists:
--> 135             a.draw(renderer)
    136     else:
    137         # Composite any adjacent images together

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     36                 renderer.start_filter()
     37 
---> 38             return draw(artist, renderer, *args, **kwargs)
     39         finally:
     40             if artist.get_agg_filter() is not None:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/axis.py in draw(self, renderer, *args, **kwargs)
   1213         # the actual bbox
   1214 
-> 1215         self._update_label_position(renderer)
   1216 
   1217         self.label.draw(renderer)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/axis.py in _update_label_position(self, renderer)
   2012         # get bounding boxes for this axis and any siblings
   2013         # that have been set by `fig.align_xlabels()`
-> 2014         bboxes, bboxes2 = self._get_tick_boxes_siblings(renderer=renderer)
   2015 
   2016         x, y = self.label.get_position()

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/axis.py in _get_tick_boxes_siblings(self, renderer)
   1996         # if we want to align labels from other axes:
   1997         for nn, axx in enumerate(grp.get_siblings(self.axes)):
-> 1998             ticks_to_draw = axx.xaxis._update_ticks()
   1999             tlb, tlb2 = axx.xaxis._get_tick_bboxes(ticks_to_draw, renderer)
   2000             bboxes.extend(tlb)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/axis.py in _update_ticks(self)
   1078         """
   1079         major_locs = self.get_majorticklocs()
-> 1080         major_labels = self.major.formatter.format_ticks(major_locs)
   1081         major_ticks = self.get_major_ticks(len(major_locs))
   1082         self.major.formatter.set_locs(major_locs)

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/ticker.py in format_ticks(self, values)
    257     def format_ticks(self, values):
    258         """Return the tick labels for all the ticks at once."""
--> 259         self.set_locs(values)
    260         return [self(value, i) for i, value in enumerate(values)]
    261 

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/ticker.py in set_locs(self, locs)
    687                 self._compute_offset()
    688             self._set_order_of_magnitude()
--> 689             self._set_format()
    690 
    691     def _compute_offset(self):

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/matplotlib/ticker.py in _set_format(self)
    789         thresh = 1e-3 * 10 ** loc_range_oom
    790         while sigfigs >= 0:
--> 791             if np.abs(locs - np.round(locs, decimals=sigfigs)).max() < thresh:
    792                 sigfigs -= 1
    793             else:

~/anaconda3/envs/tensorGpu2/lib/python3.6/site-packages/numpy/core/_methods.py in _amax(a, axis, out, keepdims, initial, where)
     28 def _amax(a, axis=None, out=None, keepdims=False,
     29           initial=_NoValue, where=True):
---> 30     return umr_maximum(a, axis, None, out, keepdims, initial, where)
     31 
     32 def _amin(a, axis=None, out=None, keepdims=False,

KeyboardInterrupt: 

def raghas(args):
    x, y = args
    z = cv2.erode((x).astype('uint8'), np.ones((3,3)))
    z = cv2.dilate((z).astype('uint8'), np.ones((3,3)))
    indcs =  np.where(z>0)
    t = np.array(indcs).T
    f = np.array([(z>0)[i-k:i+k, j-k:j+k] for i,j in zip(*indcs)])
    ff = np.array([y[i-k:i+k, j-k:j+k] for i,j in zip(*indcs)])
    f = f.reshape(*f.shape, 1)
    ff = ff.reshape(*ff.shape, 1)
    return z, indcs, (f, ff, t), z[indcs]

_ = XYZT[15]
_1 = ((_[2][m:-m,m:-m] * _[3])>.1).astype('uint8')

_1, _2, _3, *_ = raghas((_1, _[0]))

_4, _5, _6 = (_3[0]-mu1)/sig1, (_3[1]-mu2)/sig2, scaler.transform(_3[2])
_ = model.predict([_4, _5, _6])
_ = _ / _.sum(axis=1, keepdims=1)
_ = encoder.inverse_transform(_>.1)

imt = np.zeros_like(_1)
#imt[_2] = _.reshape(-1)
imt[_2] = _.reshape(-1)

plt.imshow(imt)

<matplotlib.image.AxesImage at 0x7936f03547f0>

a = np.zeros_like(imt)
for i in np.where(np.bincount(imt.reshape(-1)))[0][1:]:
    tmp = cv2.erode((imt == i).astype('uint8'), np.ones((2, 2)))
    _, cnts, _ = cv2.findContours(tmp.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
    cnts = sorted(cnts, key=lambda x: -len(x))[:1]
    tmp = np.zeros_like(tmp)
    a += cv2.drawContours(tmp, cnts, -1, int(i), -1)
    #a   += cv2.dilate(tmp, np.ones((4, 4)))*i
plt.imshow(a)

<matplotlib.image.AxesImage at 0x7936aa223b70>

_2

(array([ 39,  39,  39, ..., 215, 215, 215]),
 array([138, 139, 140, ..., 150, 151, 152]))

_4, _5, _6 = (_3[0]-mu1)/sig1, (_3[1]-mu2)/sig2, scaler.transform(_3[2])

imt = np.zeros_like(_1)
#imt[_2] = _.reshape(-1)
imt[_2] = _.reshape(-1)

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-641-cda48167d19c> in <module>
      1 imt = np.zeros_like(_1)
      2 #imt[_2] = _.reshape(-1)
----> 3 imt[_2] = _.reshape(-1)

AttributeError: 'list' object has no attribute 'reshape'

a = np.zeros_like(imt)
for i in np.where(np.bincount(imt.reshape(-1)))[0][1:]:
    tmp = cv2.erode((imt == i).astype('uint8'), np.ones((2, 2)))
    _, cnts, _ = cv2.findContours(tmp.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
    cnts = sorted(cnts, key=lambda x: -len(x))[:1]
    tmp = np.zeros_like(tmp)
    a += cv2.drawContours(tmp, cnts, -1, int(i), -1)
    #a   += cv2.dilate(tmp, np.ones((4, 4)))*i
plt.imshow(a)

<matplotlib.image.AxesImage at 0x793831965e48>

from multiprocessing import Array, Queue, Process, Pool

def mardas(args):
    i = 0
    while True:
        yield i
        i += 1
        
def gen_runner(gen, args):
    a = gen(args)
    print(a)
    while l.empty():
        x = a.__next__()
        while True:
            try:
                q.put_nowait(x)
                break
            except:
                if not l.empty():
                    return
        
def run_gen_in_par(gen, args, q, l, n_procs=4):
    ps = [Process(target=gen_runner, args=(gen, args)) for i in range(n_procs)]
    for p in ps:
        p.start()
    return ps

q = Queue(maxsize=10)
l = Queue()
ps = run_gen_in_par(mardas, (3,), q, l)

<generator object mardas at 0x7ac7bcf69a40>
<generator object mardas at 0x7ac7bcf69a40>
<generator object mardas at 0x7ac7bcf69a40>
<generator object mardas at 0x7ac7bcf69a40>

while not q.empty():
    print(q.get())

28
24
22
15
29
25
16
30
17
31
18
23

l.put(1)
for p in ps:
    p.join()

%reset_selective -f  .*(?<!_)$
from custom_imports import *
IXYZ = joblib.load('IXYZ.pkl')
results = joblib.load('11_muscle_predictions.pkl')
len(IXYZ), len(results)

(321, 167)